Sunday, May 19, 2019

2019 Senate Postcount: Main Thread

Carry-Over from 2016 Senate: Coalition 16 Labor 13 Green 3 CA 2 AC 1 PHON 1
Expected 2019: Coalition 19 Labor 13-14 Green 5-6 PHON 1 Lambie 1

Currently Coalition is likely to hold 35 seats and need two of Centre Alliance, One Nation and (Bernardi+Lambie) to pass bills.

--------------------------------------------------------------------------------------
Donations welcome!

If you find my coverage useful please consider donating to support the large amount of time I spend working on this site.  Donations can be made by the Paypal button in the sidebar or email me via the address in my profile for my account details.  Please only donate if you are sure you can afford to do so.
-------------------------------------------------------------------------------------

Welcome to my main 2019 Senate postcount thread.  This will contain outlooks for each state which I will update.  I may move any state that I do any very complex modelling on to a different thread.  In the case of Tasmania, this is only likely to happen if Lisa Singh's below the line vote starts projecting to such a level as to create a serious contest between her and Catryna Bilyk.

Some states will receive much higher detail level than others on account of the competitiveness of races.  Where races appear uncompetitive I won't be posting frequent updates.


On this page, a quota is c. 14.28%.  A candidate will be elected if they reach quota, but in the case of the last seat or two may not need to get quota to be elected.  Votes are initially counted by party (whether above the line or below) and are then gradually sorted from "unapportioned" into ticket votes and candidate votes (BTLs).  This is a long and messy process.  Initially candidate below-the-lines will be much lower than where they eventually get to so please don't say "oh so-and-so is on only 6 votes" until the unroll is finished.  As of Tuesday, the unroll for Queensland has just started.

DISCLAIMER: All assessments are provisional unless clearly stated as definite calls.  While these counts are not as difficult to project as Group Ticket counts in Victoria and WA, unexpected outcomes may nonetheless occur and I have had limited time to closely study all possible issues in the counts.

IMPORTANT: Please be aware that how-to-vote cards are weakly followed in the Senate and as a result preference flows are generally not very strong.  Do not use how-to-vote cards for modelling Senate counts unless you are familiar with the follow rates for each party's card from the 2016 election.  In most cases, a primary vote lead of 2% will not be overturned by preferences in this system.  A possible exception is in ACT.  

--------------------------------------------------------------------------------------------------------

New South Wales

(Seeking re-election: 2 Coalition 1 Labor 1 Green 1 UAP 1 Lib Dem)
Expected result: 3 Coalition 2 Labor 1 Green

Intro (Sunday):

NSW is currently 53% counted to primaries.  At a quick look the counts tend to include normal booths but often not prepoll voting centres.  These are the leaders:

Liberal/National 2.65 quotas
Labor 2.14 quotas
Greens 0.66 quotas
One Nation 0.33 quotas
Shooters 0.17 quotas
HEMP 0.15 quotas

At this stage the NSW result looks straightforward with a 4.5% gap between the leading contenders for the last two seats and anybody else, which should be much too large to bridge on preferences even if the numbers settle slightly with more counting.  

Assuming it stays like this the only interest is in the Jim Molan below the line votes as he seeks to dislodge the Nationals' Perin Davey from #3 and win the third Coalition seat.  As a BTL-only candidate Molan cannot receive above the line preferences unless Davey is excluded.  NSW doesn't typically have a lot of BTLs for other candidates and so Molan would need a large share of the Coalition's leftovers in order to exclude Davey before she could gain too much on him from other parties.  He might need something like half a quota, but I'll model this exactly if he even looks like getting half the Coalition's excess.  Across a state as large as NSW and with such a low BTL rate this seems as difficult as it always has, but I don't have any data on it.  There will be some leakage when Molan is (presumably) excluded but a lot would have to change for it to place the seat at risk.

Thursday 1:50 There are some early BTLs in which Molan is recording many times the BTL rate for both major party ticket leaders, by ratios of 8-10.  These are probably from very unrepresentative booths (haven't checked where, as it isn't easy to do) but it is interesting that he is doing this anywhere.  In 2016 both major party ticket leaders got nearly 1% on primaries but in this case they are non-incumbents and may very well get a lot less.

7:40 Some more BTLs in and the impression so far is that these are still from unrepresentative urban booths.  Molan may pick up a few percent in the end but there is no reason off these early figures to believe that he can win.

Friday 10:30: Ross Leedham who has a far better handle on this one than me is projecting Molan to land between 1 and 4% (nowhere near enough).

----------------------------------------------------------------------------------

Victoria

(Seeking re-election: 2 Coalition 2 Labor 1 Green and Derryn Hinch)
Expected result: 3 Coalition 2 Labor 1 Green 

Intro (Sunday): 

The Victorian count is relatively slow at 43% counted.  The following are leaders:

Coalition 2.38 quotas
Labor 2.27
Greens 0.82
One Nation 0.192
Hinch Justice 0.186
Labour DLP 0.175
UAP 0.162
Shooters 0.123
HEMP 0.112
Lib Dems 0.101

The first five seats are obvious but the last seat is not yet clear, as nobody fighting for it has very much.  The Coalition may move more clear of the field with the counting of prepolls and postals, or Labor's position might improve (I haven't looked at where these votes are from).  The problem for Labor on the current numbers is that the left-wing micro-party preferences will be splitting between them and the Greens (to the extent they follow ideology at all rather than spraying and exhausting) and the other problem is that there aren't all that many left-wing preferences,  though Labour DLP will help a bit.  However, if Derryn Hinch can get ahead of Labor #3 and One Nation he may be able to draw preferences from everywhere and mount some kind of challenge if the Coalition isn't too far ahead.  Hinch's vote has really collapsed since the last election and it seems absurd to even be considering his prospects when he hasn't polled 3% in a half-Senate race, but it does highlight the possibilities of the system in Victoria.  Still most likely a Coalition win at this stage.

Tuesday 6 pm: The Coalition's position has improved to 2.43 quotas.

Wednesday 6 pm: The Coalition's position has improved to 2.47 quotas.  The Coalition is now 4% ahead of Hinch which seems like it should be too much to be bridged.  Hinch could pick up an extra boost if the Labor exclusion tips the Greens over quota with a modest surplus but I am only talking hundredths of a quota here.  At the moment I think the Coalition's lead should be too much but I am yet to look in detail. Certainly the flow of Coalition preferences to Hinch (mentioned by Glenn Druery in the Age) is irrelevant.

----------------------------------------------------------------------------------------------------------

Queensland

(Seeking re-election: 2 LNP 2 ALP 1 Green + Fraser Anning)
Expected result: 2 Coalition 1 Labor + 4 way race for 3 spots
Three-way race between Greens (leading), One Nation, Labor and LNP - one to lose (probably Labor or Greens)

Queensland Senate has been moved to its own thread.

Intro (Sunday):

The Queensland count is not very advanced at 43%, but we know that the LNP thumped Labor in the House of Reps by a bruising 57-43 (compared to polls around 51-49 or 50-50).  These are the primaries:

LNP 2.55 quotas
Labor 1.65
Greens 0.82
One Nation 0.69
UAP 0.23
HEMP 0.14
KAP 0.116
AJP 0.095
FACN 0.094

At this early stage we have the Greens 0.13 quotas ahead of One Nation and Labor 0.10 quotas ahead of the LNP.  Labor are also not far behind One Nation.  Whoever is last out of these four after preferences loses, but the LNP (and also One Nation) should do well out of UAP who are the next biggest source of goodies, Clive Palmer's tilt having clearly failed.  The 2PP from the Reps implies that these primaries might be understating the conservatives at this early stage, and bear in mind that certain North Queensland seats (especially Flynn) tend to have big turnarounds in the LNP's favour on postals.  It is also especially notable that at this stage the urban seats are slightly over-represented in the count compared to the far-northern regionals, so the count appears skewed to the left (so Labor's miserable 24% primary might be an overestimate!)

So it may well be that Labor is in trouble here, but we'll have to see where the parties stand when the primary count is more complete, as maybe the urban skew issue will affect the Greens more than them.  One Nation's Malcolm Roberts is going relatively well so far and based on the skew issue might do even better.

Tuesday 5:30 pm: Ross Leedham has been unskewing the primary vote count and has recently posted this projection for the final seat (my quota numbers in brackets afterwards)

Greens - 10.10% (.707)
ON - 9.97% (.698)
ALP - 9.11% (.638)
LNP - 8.87% (.621)

It's possible we won't know the loser of this race for a long time but given the right-wing skew of the preference base it could be that One Nation and the LNP might both win leaving Labor vs Greens for the final seat (and it's not impossible Labor could catch up there.)  The projection does not account for declaration votes.  In total right wing micros have about 11% vs about 4% for left micros, so it is hard to see the left parties both getting up at this stage.

Friday 8:00 -  In the live count the LNP now has 2.76 quotas, One Nation 0.71, Greens 0.69, Labor 1.61.  The Labor vs Greens race is impossible to call and likely to remain that way until the button.  I am hoping to look at the patterns from last time to get some sort of handle on it, but this involves complex modelling, and my feeling is the preference flow to Labor from KAP etc will weaken.  One Nation are not necessarily safe either as there is no love lost between them and FACN and it's possible the LNP will get more conservative preferences than they will.  But given the overall right-wing skew of the preference pool it's rather difficult to see them missing out.  Most likely Anning's voters will preference One Nation anyway, and really the only risk is that other parties preference One Nation much more weakly on account of Pauline Hanson herself not being on the ballot.  I think both conservative parties should win easily.

Saturday 5:00: Ross Leedham has been tweeting some unflattering figures for the Greens concerning their performance in PPVCs, of which significant numbers remain uncounted.

Sunday: Moved to own thread.
----------------------------------------------------------------------------------------------------------

Western Australia

(Seeking re-election: 2 Coalition 2 Labor 1 Green 1 One Nation)
Expected result: 3 Coalition 2 Labor 1 Green

Western Australia is at about 52% counted and these are the leaders:

Liberal 2.82 quotas
Labor 1.98
Green 0.89
One Nation 0.37

There's no point mentioning the rest because the first three are so far ahead that I can't see further counting bringing One Nation into the mix.  I don't expect to be commenting further on WA.

------------------------------------------------------------------------------------------------------------

South Australia

(Seeking re-election: 3 Liberal, 1 Labor, 1 Green, Retired at election: Tim Storer)
Expected result: 3 Coalition 2 Labor 1 Green

South Australia is about 57% counted and here are the leaders:

Liberal 2.54 quotas
Labor 2.20
Green 0.83
One Nation 0.32
UAP 0.20
Centre Alliance 0.18
HEMP 0.15
Animal Justice 0.12
Conservatives 0.11

At the moment the Liberals are over 3% ahead of One Nation, a large lead that is unlikely to be caught based on either remaining primary votes or preferences.  The Centre Alliance has bombed out completely, suggesting its current Senators will struggle to get elected in 2022.  Labor also doesn't seem to be a threat as the usual issue occurs here of the left votes splitting two ways between them and the Greens.  I will keep an eye on this one but don't expect the outcome to change.

------------------------------------------------------------------------------------------------------------

Tasmania

(Seeking re-election: 3 Labor, 1 Liberal, 1 National, 1 Green)
Expected result: 2 Liberal, 2 Labor, 1 Green, Jacqui Lambie
Identity of second Labor winner to be confirmed, probably Catryna Bilyk
Lisa Singh has polled a large below the line vote but does not appear to be winning

Tasmania is at 63% counted and here are the leaders:

Labor 2.17 quotas
Liberal 2.17
Green 0.89
Jacqui Lambie Network 0.60
One Nation 0.25
United Australia 0.18
Shooters 0.12

There are also two indies with about 1% (0.07 Q) apiece - Craig Garland running as a grouped independent and Steve Mav running as an ungrouped BTL-only indie.  (Although Mav has lost his deposit, his 1.04% is still about 25 times better than any other ungrouped-column candidate in Australia.)  Steve Martin's Nationals attempt has also bombed out getting just more than these two. Lambie's vote has held up remarkably well despite a lack of resources and so much competition from similar parties, and indeed she seems to have taken a few votes from the majors.

The seat breakdown by party here is not going to change.  There is, however, the issue of Lisa Singh's below-the-line vote.  Singh was demoted to what is wrongly called an "unwinnable" position in 2016 and won from it, getting 6.1% of the vote.

In this case it might seem that half Labor's excess over one quota (so that would be 8.4%) would help Singh beat #2 Catryna Bilyk or perhaps Lambie.  But this is not the case, because from that point Singh could only attract below the line preferences until such time as Bilyk was eliminated, while Bilyk would keep getting above the line preferences from excluded micro-parties.  We saw this in the 2016 count where Bilyk had gained 1.6% more than Singh at the time when Singh was elected.  A further disadvantage for Singh is that the available preference pool for parties with an ATL box is slightly larger this time than in 2018 (I get 18% compared to 14.4%).  The target for Singh to win seems to be a primary of approaching 9.5%, about 1.5 times her 2016 vote.  I have seen samples from four booths that showed that she was getting something at least comparable to last time, and much higher in one small booth, but as the primary vote unrolls we will get a better idea.  In the seemingly unlikely case that Singh is competitive I'll transfer this to a new thread.  My suspicion at the moment is she'll finish seventh, as she would have in 2016 had that been a half-Senate race.

Wednesday 5:15 Unrolls of some early booths have started.  These are too small to draw reliable conclusions from as two are hospital booths that may have changed their list of hospitals.

Thursday 2:40: Still only a very minimal unroll of unrepresentative booths, with under 1000 votes unrolled.  I also have the unroll from a single booth from scrutineering.  At this stage I am tracking both average swing by polling place on the Singh vote and total swing from votes counted, and waiting until the sample becomes significant.

Thursday 7:40: Two significant booths in with a zero swing on the Kingston PPVC booth (Franklin) and 1 point against Singh in Goodwood (Clark).  Not an encouraging start at reaching a very difficult target.

Friday 7:50:  I have 4.6% of the count unrolled at matching booths so far.  Some large prepolls were added in Braddon today.  I have Singh tracking for the 6-7% range at the moment but this is still very rubbery.

Monday 2:30: Off 7.3% of the count unrolled at matching booths, I have Singh tracking for around 6.4%.

Monday 10:40: Off 13.7% I still have Singh tracking for around 6.4%.  The count is, however, dominated by Bass votes with no Lyons votes included at all.  I'm using the average of two projections to project at the moment (swing by booth and swing by total votes) and the latter gives a very slightly higher figure than the former.  I'll switch to using wholly the latter once the range of booths is more representative.

Tuesday: Off 26.7% I have Singh tracking for around 6.1%.  The count now has some Lyons votes in it.  It is still skewed to areas that didn't vote strongly for Singh last time.

Wednesday: I added just Franklin booths for today so far and the projection for Singh dropped to 5.8% with minor swings against in many of her better Franklin booths.  It is all academic now but it will be interesting to see if this is also the case in the Hobart City booths in Clark.  If so Singh's eventual vote might drop down to the low 5s.

Friday: With 43.8% booth-matched I have Singh tracking for 5.9% but I am still lacking a significant sample from Clark booths.

Sunday: With 47% booth-matched including the Clark booths added over the weekend I now have Singh tracking for 5.7%.

Thursday: With around 70% booth-matched including more Clark booths I now have Singh tracking for 5.5% but we will know exactly what she has soon enough anyway (she has just passed 5% in the live count.)  As the graphs below (based on booths with at least 200 votes) show there is a clear pattern that Singh does better relative to her 2016 vote in booths where she recorded few votes last time.




Tuesday: Singh has finished on 5.7%.
---------------------------------------------------------------------------------------------------------

ACT

(Seeking re-election: 1 Labor 1 Liberal)
Expected result: 1 Labor 1 Liberal (CALLED)

A quota in the ACT is a third (33.33%).  The ACT primary count is at 51% complete with the following scores:

Labor 1.190 quotas
Liberal 0.896
Greens 0.588
Pesec (IND) 0.158
UAP 0.062
SUS-A 0.048
FACN 0.027
CDP (ungrouped) 0.014
ungrouped inds 0.017

The out-there scenario that has been considered for the ACT Senate from time to time is the Greens taking the Liberals' second spot.  At present they are .308 Q behind with .348 Q in spare Labor and Pesec preferences.  But for the Greens to pass Zed Seselja off these preferences, the Greens would need to gain at .88 votes/vote, when in 2016 they only gained at .54 votes/vote from Labor.  The 2019 ballot paper will have slightly lower exhaust, and Labor votes might flow more strongly anyway, but even so a .88 flow off Labor is currently impossible. Even if it were possible the UAP, FACN and CDP preferences would help Seselja.

I also don't have any data on flows from Pesec but I would doubt his preferences would flow more strongly to the Greens than Labor's, or at least not much more.  So I think the Greens are at least 0.1Q (3.3%) behind where they need to be relative to the Liberals at the moment and need to make massive improvements on remaining primaries before this could be taken seriously.  It may be that the Liberals will improve instead on postals.

Thursday: The Liberals have now improved to very close to a quota with their lead out to nearly .44 quotas, though this may come back slightly.  I have had scrutineering reports that while the Labor flow to the Greens has increased slightly, the flow from Pesec is only slightly advantageous.  On either of these bases I would be happy to call it, let alone both.

---------------------------------------------------------------------------------------------------------

NT

(Seeking re-election: 1 Country Liberal 1 Labor)
Expected result: 1 CLP 1 Labor (CALLED)

A quota in the NT is one third (33.33%).  The NT has a slow count with 31.5% counted and the CLP are on 1.12 quotas, Labor 1.08, Greens 0.36 and the rest no more than 0.14 apiece.  Even if one major falls below a quota the others cannot catch them.  I won't be updating NT.

23 comments:

  1. Kevin, I notice on the AEC's "First preferences by candidate" pages that for each party they have a line called "Ticket votes", then lines for the candidates with no votes showing yet, and a line called "Unapportioned". All just like last time, as far as I recall - except in each case there is a largish number opposite the unapportioned and a much smaller number opposite Ticket votes. I think (a.f.a.i.r.) last time "unapportioned" meant the total of the BTLs so it should be the smaller number. If they're lumping everything into Unapp'd until examined more closely, then why are there very small numbers creeping into "ticket" while absolutely none get into the individual lines? Or have they swapped the entries, and the small numbers represent the total BTLs (they're perhaps *too* small for that to be the explanation)? Truly they are not very good at explaining what they do - they need a plain language consultant to work with their web designers!

    ReplyDelete
    Replies
    1. In 2016 I wrote this: "However some booths temporarily appear in the results in a partly separated-out form, so that the ticket votes for the candidates are shown but the below-the-line votes are not split up by candidate. " The explanation of what was going on was a nightmare last time and I expect no better this time.

      Delete
    2. A matter begging for a submission to JSCEM. I might draft one and send it to you for co-signing.

      Delete
  2. What are some of the impediments that stops a 'Grouped' Senate candidate from setting up their own political party before an election? I am sure Craig Garland would have attracted a competitive number of votes if he was running under a Party banner on the Senate paper, rather than a 'Group O' header? Most people just scan things pretty quickly when they vote and I feel he wasn't recognised by a lot of people that may have otherwise ended up voting for him.

    ReplyDelete
    Replies
    1. Need 500 signatures who will confirm that they are actually members of a party when contacted (a small tolerance for failures is allowed) and to apply in sufficient time for registration process to occur.

      I think Garland just didn't get as much media cover as in the Braddon by-election with so much else going on in the campaign.

      Delete
  3. Kevin, any word on the Campaign to put Steve Mav 44th ? I saw a FB page that people were sending photos of ballots to and there were clearly quite a few. Sadly the AEC doesn't list 44 votes, but I'd be keen to know if their was any traction with this beyond the few hundred people on a FB group.

    ReplyDelete
    Replies
    1. This won't be known until the AEC publishes the data files after the button press in several weeks.

      Delete
  4. Given the Nationals' reaction to the Molan campaign we could see some BTLs for Davey as well. If either of those sets of BTLs leak significantly to ONP things could get interesting with the large SFF vote.

    ReplyDelete
  5. So in Qld there is more than a quota beneath Palmer... You rate KAP as a chance but surely at those low counts a few good preference flows could put Anning or Bernardi ahead of KAP...

    ReplyDelete
    Replies
    1. I don't rate KAP as a chance. Or UAP for that matter. Only Coalition, Labor, Greens, One Nation - three of these four will win the last three seats. I just included them in the list to show the major preference sources.

      Delete
    2. Preferences no longer flow. They sputter.

      Delete
  6. Hi Kevin. I like your work. The Senate exhaust rate for votes from excluded groups could be fairly high and have a significant impact on tighter contests such as the QLD Senate. Would you give your thoughts on estimating the exhaust rate.

    ReplyDelete
    Replies
    1. Sorry for being slow replying to this comment. The real exhaust rates in 2016 were not as high as some people have claimed, and the reason for this is that many of the exhausted votes left the count after the results were no longer in doubt. The AEC preference throws go beyond the point where the result is determined.

      Effective exhaust in Queensland in 2016 was 4.2% and that was for a race that ended up as a contest between One Nation and micros. This race will end up as a contest between the four biggest preference attractors in the field. It might end up with two crossing quota and a race between the final two (which currently looks like Labor and Green). A vote will only exhaust at full value if it hits none of the big four, and those votes exhausting at part value will only go to exhaust at say 5% of value. I expect the effective exhaust in the Queensland race to be lower than last time, but it still will slightly hinder the party that is last and trying to catch up.

      Delete
  7. In Victoria the coalition is now sitting on 2.483 of a quota. 4.1% ahead of Hinch. Druery says Hinch is getting preferences from everywhere. But under this new system everyone gets preferences from everywhere. The Liberals will be getting preferences too. To make up 4.1% of the vote is impossible, particularly when you primary vote was so low. Druery is dreaming.

    ReplyDelete
    Replies
    1. Tend to agree, 4% looks like too much. Druery also says Hinch is getting prefs from the Liberals but so what, those will never be distributed.

      Delete
    2. Thursday 8.30 pm update. Libs out to 2.50 of a quota and extending the margin to 4.36% over Hinch. 30% of the vote to come which will favour the Libs. And Hinch still trails Labor. And one would expect a lot of Labour DLP votes are actually votes for Labor given their position on the ballot paper. So Labor should get preferences to stay ahead of Hinch.

      Delete
  8. In regards to the Qld contest, is it reasonable to do a rough guess of the potential preferences flows to Greens versus Labor based on the 2016 count, or were there still too many undistributed preferences left by the time the final Green candidate was removed from the count (by virtue of being elected)?

    Or could people recrunch all the numbers from the date entry of the 2016 count to get a better guesstimate on likely Labor vs Green preference flows (noting that some of the parties are different this time).

    ReplyDelete
    Replies
    1. I think the best method is to download the individual votes and find and sort by party those votes for specific micros that went to Labor or the Greens ahead of LNP or ONP. But I also suspect that this will be unreliable because the flow to Labor of, eg, Katter preferences will almost certainly weaken. It is a big exercise but I will probably do it when the pressure of other things recedes, maybe over the weekend, if nobody else has done it.

      Delete
    2. Does any of the scrutineering indicate how the preferences are flowing from UAP. We know they were directing number 2 preference to the Liberals. Just keen to know if this is happening. Particularly in Queensland. Interested to know flow rate to Libs vs One Nation vs Labor vs Greens.

      Delete
    3. I haven't seen any scrutineering reports on preferences from any state.

      Delete
  9. With regard to PHON not being safe in Queensland you should remember how Malcolm Roberts was able to rise from 0.1941 of a quota to 0.7764 of a quota at the time the Lib candidate was elected. The Lib candidate started with 0.5851 of a quota. With regard to Fraser Annings voters are you seriously suggesting they would not mark PHON as one of their preferences? Or put Lib, Lab or Green higher? I think you will find most will put PHON 2nd.

    ReplyDelete
    Replies
    1. Suspect so too - I expect almost no FACN voters will follow their card (if it was even handed out at all) and that a lot will preference One Nation irrespective of arguments between them. But we'll see.

      Possible that Hanson not herself being on the ballot will weaken the preference flow to ON, but that's also not likely to make that much difference. The LNP may get a better flow off UAP, but UAP voters and PHON voters have a lot of crossover so I wouldn't be surprised to see a lot of them preferencing PHON as well.

      Delete
  10. Thanks for your responses Kevin. One could argue Malcolm Roberts has made himself almost as well known in Queensland as Pauline. Same could be said for Steve Dickson. Hah hah. With regards to preference flow from UAP at position C, ON was to the left at B and LN to the right at D.

    On Victoria Hinch is now trailing Libs by 4.8%.

    ReplyDelete